Automatically create dataset JSON schemas through introspection

[chrisschopp]

Description

When a new dataset is added to kedro-datasets, the JSON schema that enables Kedro catalog validation in an IDE must be manually updated.

These schemas could be created programmatically through introspection (as suggested by @datajoely).

Context

Manually creating a JSON schemas is an additional maintenance step (that may be easy to overlook).

Possible Implementation

1. The inspect module could be used to get the dataset's attributes.
2. A pre-commit hook could generate the required kedro-catalog-X.XX.json (only for the current package version).
   a. When a developer commits a new dataset, the JSON schema is created for them.
3. Remove references to updating jsonschema from CONTRIBUTING.md and PULL_REQUEST_TEMPLATE.md.

Possible Alternatives

The introspection code could be triggered manually instead of by a pre-commit hook. The hook would fully automate it, though.

[astrojuanlu]

This would be nice to have for the Kedro team, but it's not something we'll be doing in the short term. If someone wants to open a PR, go for it!